基于多标记图分割的遮挡下多目标分割及跟踪算法

针对遮挡下的多目标分割及跟踪问题,提出一种新的方法,其将空间-彩色混合高斯模型融入到能量函数最小化框架中。当遮挡没有发生时,为每个目标分别构建一个空间-彩色混合高斯形状描述。一旦多目标相互遮挡发生,一个多标记能量函数将在相互遮挡区域建立起来,而后利用多标记图分割技术将其最小化,从而实现对遮挡过程中所有目标进行同时标记和定位。此外,利用多个视频序列证实了所提出算法的性能。

A Multi-label Graph Cut-based Algorithm for Multi-object Segmentation and Tracking under Occlusion

A new method is proposed for segmenting and tracking multiple objects through occlusion by integrating spatial-color Gaussian mixture model (SCGMM) into an energy function minimization framework. When occlusion does not occur, a SCGMM is learned for each object. When the objects are subject to occlusion, a multi-label energy function is formulated building on the learned SCGMMs, and then minimized using the multi-label graph cut algorithm, thus leading to both the segmentation and tracking results of the objects with occlusion. Experimental validation of the proposed method is performed and presented on several video sequences.